Analysis of Tactile, and Audio-Tactile Display-Based Regulatory Human-in-the-Loop Control Systems

Abstract

Throughout history a great variety of tasks has involved the utilization of a combination of mechanical or electromechanical equipment and a human operator. These tasks range from simple activities such as using a hammer to install a nail to complicated maneuvers executed by fighter plane pilots. The inherent interdependence of humans and machines forms a closed loop control system known as the human-in-the-loop control system. The objective of the system is to complete a task within certain limitations, such as time, imposed by operating requirements. In modern times, artificially created guidance signals provide feedback to the human user to guide him or her in executing the desired task in an optimal manner. The most frequently used guidance is based on visual cues such as lights and monitors. Its role has been studies by a great many researchers as its applicability is trivial due to its intuitive nature. In certain situations however, visual guidance cannot be provided and humans may need to rely on proprioception, auditory or tactile guidance. While application examples of tactile and audio-tactile support have been known, their optimal selection of operating parameters and humans’ behavior in response to receiving these signals have been disproportionately understudied. Current dissertation aims at identifying the parameters and efficiency of vibrotactile and audio-vibrotactile feedback signals in their ability to provide guidance in human-in-the-loop control systems. This work also sets out to describe human behavior in response to these guiding signals. In possession of the understanding of human behavior, development of an active controller is targeted with the goal to further optimize human performance. The Chapters 1 and 2 of the dissertation review the literature related to human-in-the-loop control systems and the various guiding signals used in these systems. Chapter 3 presents the research questions and hypotheses, while Chapter 4 outlines the necessary experiments to analyze the research questions. Chapter 5 presents a custom made, state of the art device, its utilization throughout various 1 and 3 dimensional target seeking experiments is discussed through Chapters 6-9. The setup, results and discussion of the experiments lead to conclusions regarding the fundamental research questions. Final conclusions are summarized in Chapter 10. A review of potential application examples is included in Chapter 11 while a proposal for future research is presented in Chapter 12. Analysis of experiment results reveals that the best performing displays utilize audio-tactile technology, where the indication of reaching the target position is signaled by an auditory beep. The seconds best performing group of displays provide tactile-only guidance, wherein the on-target position is indicated by the lack of tactile feedback. The bottom of the ranking list is populated by tactile-only displays which provide no differentiated signal when the subject reaches target. These results apply to both 1 and 3 signal channel scenarios. System identification based transfer function model generation showed greatly limited applicability to successfully and universally describing human behavior. A custom made system of behavior primitives was introduced which describes the digitized version of the recorded human behavior in smaller building blocks. Main findings indicate that people utilize a combination of oscillatory movements, methodical step-like movements and small-sloped “creeping” towards the desired target. Stemming from the identified motion primitives, a non-linear active controller was developed which resulted in a significant improvement is the operators’ ability to successfully complete the presented task.